NICE Recommendations for Hyperglycaemia in ACS
The GDG considered two related review/clinical questions regarding the initial inpatient management of hyperglycaemia in patients with ACS:
To answer the first question for patients with prior diabetes, the GDG reviewed data from the DIGAMI, DIGAMI-2 and HI-5 trials, which were the only randomised studies identified that fulfilled the predefined criteria of targeting blood glucose control in ACS patients with documentation of blood glucose levels at baseline, or provided a definition of hyperglycaemia. Intensive insulin therapy was defined as an intravenous infusion of insulin and glucose with or without potassium. While the original DIGAMI study showed a well-known reduction in mortality with intensive insulin versus conventional therapy that has driven much clinical practice ever since, ACS treatment has evolved considerably since 1995 with widespread and routine use of therapies including statins, antiplatelet agents and revascularisation, all of which may abrogate the potential benefits of intensive insulin therapy and explain why the mortality benefit was not replicated in the more recently published DIGAMI-2 or HI-5 studies. Unsurprisingly, when all three trials were combined in the GRADE meta-analysis, no mortality benefit was seen. The evidence statement reads, 'very low-quality evidence from three studies, with a total of 1640 patients, showed that intensive insulin did not significantly reduce overall mortality compared with standard care after a follow-up of up to 3.4 years (RR 1.03, 95% CI 0.65 to 1.62).' All 10 evidence statements derived from the GRADE meta-analysis were considered to be of very low-quality evidence on the basis of downgrading for inconsistency (non-contemporary, non-UK studies), indirectness (inclusion of non-diabetic patients) and imprecision (wide CIs).The strongest treatment effect identified was actually for potentially harmful hypoglycaemia in the intensive insulin-treated subjects: 'very low-quality evidence from two studies, with a total of 1401 patients, showed that hypoglycaemic events were significantly more likely in the intensive insulin group than in the standard care group during the initial 24 h of treatment (RR 19.32, 95% CI 5.79 to 64.41)'. Some very low-quality evidence in selected subgroups (low-risk patients without previous insulin therapy, and patients in whom a blood glucose level of ≤8 mmol/l was achieved) did show a reduction in short- and long-term mortality, but the GDG felt that the very low quality of this evidence in small patient numbers was not reliable enough on which to base firm treatment recommendations.
For the second review/clinical question in patients without prior diabetes, the GDG considered two randomised trials, and one observational study, which was included as it was a large UK-based study looking specifically at patients with ACS and hyperglycaemia who had no previous diagnosis of diabetes. The DIGAMI studies were not considered for this review/clinical question, as they predominantly enrolled patients with a prior diagnosis of diabetes. As for the analysis of patients with prior diabetes, the GRADE meta-analysis for this research question produced 10 evidence statements of very low or low-quality evidence which showed no convincing benefit of intensive insulin therapy. The observational data from MINAP did show a mortality benefit in patients treated with intensive insulin therapy, but the GDG could only give minimal weighting to this very low-quality evidence, particularly when the overall balance lies with no overt benefit.
Overall, the level of evidence available allowed the GDG to draw up a single recommendation covering patients with and without prior diabetes:
Do not routinely offer intensive insulin therapy to manage hyperglycaemia in patients admitted to hospital for an ACS unless clinically indicated (see summary of recommendations Box 1 )
However, the GDG acknowledged that there were potential well-established hazards to not treating hyperglycaemia per se in any acutely unwell patient, such as ketoacidosis or hyper-osmolar syndrome, and so a further recommendation was also included:
Manage hyperglycaemia in patients admitted to hospital for ACS by keeping blood glucose levels below 11.0 mmol/litre while avoiding hypoglycaemia. In the first instance, consider a dose-adjusted insulin infusion with regular monitoring of blood glucose levels.
From Evidence to Recommendations: In-patient Metabolic Management of Hyperglycaemia
The GDG considered two related review/clinical questions regarding the initial inpatient management of hyperglycaemia in patients with ACS:
What is the optimal inpatient metabolic management of hyperglycaemia in a person presenting with ACS and hyperglycaemia; and who also has or does not have a previous diagnosis of diabetes mellitus?
To answer the first question for patients with prior diabetes, the GDG reviewed data from the DIGAMI, DIGAMI-2 and HI-5 trials, which were the only randomised studies identified that fulfilled the predefined criteria of targeting blood glucose control in ACS patients with documentation of blood glucose levels at baseline, or provided a definition of hyperglycaemia. Intensive insulin therapy was defined as an intravenous infusion of insulin and glucose with or without potassium. While the original DIGAMI study showed a well-known reduction in mortality with intensive insulin versus conventional therapy that has driven much clinical practice ever since, ACS treatment has evolved considerably since 1995 with widespread and routine use of therapies including statins, antiplatelet agents and revascularisation, all of which may abrogate the potential benefits of intensive insulin therapy and explain why the mortality benefit was not replicated in the more recently published DIGAMI-2 or HI-5 studies. Unsurprisingly, when all three trials were combined in the GRADE meta-analysis, no mortality benefit was seen. The evidence statement reads, 'very low-quality evidence from three studies, with a total of 1640 patients, showed that intensive insulin did not significantly reduce overall mortality compared with standard care after a follow-up of up to 3.4 years (RR 1.03, 95% CI 0.65 to 1.62).' All 10 evidence statements derived from the GRADE meta-analysis were considered to be of very low-quality evidence on the basis of downgrading for inconsistency (non-contemporary, non-UK studies), indirectness (inclusion of non-diabetic patients) and imprecision (wide CIs).The strongest treatment effect identified was actually for potentially harmful hypoglycaemia in the intensive insulin-treated subjects: 'very low-quality evidence from two studies, with a total of 1401 patients, showed that hypoglycaemic events were significantly more likely in the intensive insulin group than in the standard care group during the initial 24 h of treatment (RR 19.32, 95% CI 5.79 to 64.41)'. Some very low-quality evidence in selected subgroups (low-risk patients without previous insulin therapy, and patients in whom a blood glucose level of ≤8 mmol/l was achieved) did show a reduction in short- and long-term mortality, but the GDG felt that the very low quality of this evidence in small patient numbers was not reliable enough on which to base firm treatment recommendations.
For the second review/clinical question in patients without prior diabetes, the GDG considered two randomised trials, and one observational study, which was included as it was a large UK-based study looking specifically at patients with ACS and hyperglycaemia who had no previous diagnosis of diabetes. The DIGAMI studies were not considered for this review/clinical question, as they predominantly enrolled patients with a prior diagnosis of diabetes. As for the analysis of patients with prior diabetes, the GRADE meta-analysis for this research question produced 10 evidence statements of very low or low-quality evidence which showed no convincing benefit of intensive insulin therapy. The observational data from MINAP did show a mortality benefit in patients treated with intensive insulin therapy, but the GDG could only give minimal weighting to this very low-quality evidence, particularly when the overall balance lies with no overt benefit.
Overall, the level of evidence available allowed the GDG to draw up a single recommendation covering patients with and without prior diabetes:
Do not routinely offer intensive insulin therapy to manage hyperglycaemia in patients admitted to hospital for an ACS unless clinically indicated (see summary of recommendations Box 1 )
However, the GDG acknowledged that there were potential well-established hazards to not treating hyperglycaemia per se in any acutely unwell patient, such as ketoacidosis or hyper-osmolar syndrome, and so a further recommendation was also included:
Manage hyperglycaemia in patients admitted to hospital for ACS by keeping blood glucose levels below 11.0 mmol/litre while avoiding hypoglycaemia. In the first instance, consider a dose-adjusted insulin infusion with regular monitoring of blood glucose levels.